JP4615776B2 - Edge enhancement processing processor and method by adjustable threshold setting - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates generally to digital image processing, and more particularly to edge features of a printed or displayed image generated from a low resolution image file that includes binary image data or a mixture of binary image data and grayscale image data. The present invention relates to a digital image processing system and a method for improving the performance.
[0002]
[Prior art and problems to be solved by the invention]
Jagged edges and lines are a common problem with printing low resolution binary image data. Many efforts have been made to reduce or eliminate jagged lines. In this regard, see US Pat. No. 6,021,256. In the above specification, an edge of a mixed image file of a binary image and a gray level image using a binarization and sorting unit to screen data representing binary data from the mixed image input file of the binary image and the gray level image. A system for improving the system is disclosed. Next, the sorted binary data is supplied to a binary data edge enhancement processor for edge enhancement processing. The output from the binary data edge enhancement processor unit is supplied to the data mixing unit along with the original image data. The data mixing unit determines whether the original image data is part of the grayscale image. If the data mixing unit determines that the original data is binary image data, the output from the binary edge enhancement processor unit is supplied as the output of the processing system. The problem that the system described in the specification operates normally while the image data represents color separation image data that has undergone undercolor removal and / or gray component replacement and / or color conversion processing; and Problems can arise with the generated image data having grayscale values that do not reach the values expected by an edge enhancement processor configured to operate on the binary image data. Therefore, the edge enhancement processor considers all the image data as gray level image data, and at least a part of the image data is preferably selected after the edge enhancement processing than the image data input to the edge enhancement processor. Output image data such as data is selected.
[0003]
The above-mentioned requirements and objectives and other requirements and objectives are achieved by the present invention described below.
[0004]
[Means for Solving the Problems]
According to a first aspect of the invention, an edge enhancement processing system is provided for modifying image data at a predetermined pixel location to include grayscale image data to reduce jaggedness in the image. The system includes an adjustable threshold device that determines a current binary pixel value for a current input gray level pixel according to a thresholding criterion; and for adjusting the threshold within the threshold criterion An operator with input access to the threshold device; and a current binary pixel according to a predetermined criterion for determining to modify the current pixel to a grayscale value to reduce jagged edges of the image An edge-enhanced image processing device that inspects each adjacent binary pixel.
[0005]
According to a second aspect of the present invention, an edge enhancement method for processing image data is provided. The method includes determining an adjustable threshold within a threshold criterion in response to input from an operator; according to a threshold criterion that uses the threshold; Determining a binary pixel value; according to a predetermined criterion for determining to modify the current binary pixel to a grayscale value to reduce jagged edges of the image; Inspecting each pixel adjacent to the image; using gray scale values instead of current binary pixels to reduce jagged edges of the image.
[0006]
According to a third aspect of the present invention, an edge enhancement method for processing image data is provided. The method includes processing image data using undercolor removal and / or gray component replacement; whether undercolor removal and / or gray component replacement is used (or a range of such use) And adjusting the edge enhancement processing of the image data.
[0007]
The invention and further objects and advantages of the invention will become more apparent after reading the following detailed description of the preferred embodiment.
[0008]
DETAILED DESCRIPTION OF THE INVENTION
In the detailed description of the preferred embodiments of the invention presented below, reference is made to the accompanying drawings.
Although the method described herein is for a full process color system that performs multiple color separation, the described invention is not limited to black and white systems and accent color systems. ) Is also possible. Here, for purposes of clarity, a method and apparatus for processing image data for only one of each color separation of a multi-color separation image forming system is described. The extension to all color separations is self-evident, such as by having an additional or parallel system for each color, or by processing different colors sequentially. The input image to the system becomes continuous-tone color separation (post-RIP rasterized image) after GCR (Gray Component Replacement) and UCR (Under Color Removal) processing has already been performed. It seems that there is. The input image data is gray level image data extracted by scanning a document with a scanner. FIG. 1 shows a schematic diagram of an image processing system according to the present invention. During real-time operation of the printing system, 1-D so that a customer who desires personal or preference at the last moment can color-separate an already rastered image (such as redder, greener, etc.) A (one-dimensional) look-up table (LUT) 12 or global color process control (reprogrammable) is used to act on the input data. Next, the input data modified as output data by the LUT 12 is input to the adaptive screen analysis unit 14 (image segmentation) and analyzed to generate an image type ID function (in this case, contrast index). To do. This contrast index acts as a pointer for obtaining the mixing factor (BC1, BC2) of the selected halftone screen. In this embodiment, it is assumed that only two screens are used simultaneously (text screen and image screen). It can also be understood that an intermediate screen can be used. See US Pat. No. 5,956,157 for this (the present invention includes the contents of that specification). In this embodiment, the contrast index is one known method that uses the basic concept of image segmentation and a fuzzy logic approach to allocate the percent used for each selected screen.
[0009]
In addition, the modified input contone data is passed through two screeners or LUTs 18, 20 for halftoning simultaneously. In each screener LUT block, the input continuous tone data is halftoned only by a screener (such as a textual ultra-high frequency soft screen) under the control of a screening address calculator having inputs from the pixel clock and line clock. The halftone correction value is the output from each screening block 18, 20. In the case of a rational screen, the repeatedly calculated address of each halftone block for the two screen selections is not necessarily the same. Next, a blending operation is performed in the processor 24 taking into account the blending factor and halftone value of the entire screen, thereby outputting a modified halftone value (mixed halftone value) based on the result. The Each edge of unsaturated text / graphics is mainly likely to use a high frequency soft image screen (which uses partial dots in the growth pattern), while the interior of relatively large text is primarily low frequency (mixed dots (Growth pattern) Since the possibility of using a screen is even higher, fine details are retained and at the same time EP stability in a large area is achieved. Furthermore, since each edge of unsaturated text uses a high frequency screen, the unsaturated text is not degraded by normal low frequency screen processing (anti-aliasing effect is roughly equivalent to that made for unsaturated text and graphics). ). Mixing each screen also reduces artifacts at the boundaries of each image type. This reduces the moire problem caused by scanning an input image having high frequency features and outputs a halftone screen with a fixed screen (screen angle, screen frequency).
[0010]
As represented in US Pat. No. 5,694,224 for gray level printing, each pixel has the potential to be modified to several different dot sizes or densities (and thus different gray levels). There are at least three gray levels. Here, in the binary system, two levels (background and maximum density) can be taken. However, instead of simply providing an independent gray level for each pixel, multiple pixels may be organized together to form a superpixel or cell. Next, each pixel in the cell has a gray level. The human visual response integrates the various gray levels of individual pixels in a cell into one perceived gray level for the cell. This is the same as the basic concept of binary halftoning. However, the number of tone skills for a cell is greatly increased since a different number of gray levels is available for each pixel. For example, instead of having only two levels in binary halftoning for each pixel, it is possible to have 256 levels (including zero) by gray level printing for each pixel in the cell. In order to obtain different desired results, it is possible to form dots for each pixel of the cell in a different number of ways. The dots can be formed as “full” dots, “partial” dots, “mixed” dots, or fixed dots to provide gray level halftoning. The partial dot formation step and the confusion dot formation step are described in the aforementioned US Pat. No. 5,694,224.
[0011]
To this extent, the system produces an anti-aliasing effect on non-saturated text and reduces the moire caused by the output screen's beat on input, while at the same time being stable to the EP process. Can be maintained. The system also needs to produce an anti-aliasing effect on unsaturated text. In addition, for color systems, GCR and UCR are often used, so some of the originally saturated text (in monochromatic) has changed to near-saturated text. To solve this problem, a programmable adjustable threshold / detector 26 is used for the mixed halftoned data (see FIG. 1). Thus, a mixed gray level halftone value that exceeds a certain threshold is converted to a binary 1 value by the GRET adjustable threshold / detector 26 and the remainder is input to the GRET antialiasing detector 28. Before, it is set to a binary zero value. In this regard, see US Pat. No. 5,450,531 and US Pat. No. 5,600,761. While other gray level edge enhancement processors are used to enhance each edge of saturated text, the present invention includes the contents of the aforementioned US patent document by reference to the disclosure of GRET processing. In the proposal of the antialiasing edge improvement output by GRET, a pointer is set in the LUT including the multi-level output value so that each edge is smooth. As shown in FIG. 1, each intensity LUT (gray value) can be provided for more or less smooth or line width control. This special GRET intensity has been selected by input to the LUT 30. Of course, based on the GRET algorithm for multilevel images, the detector 26 may or may not have any non-binary values (high but not saturated and / or low values) present in the examination window. Judgment is also made. If there are other gray values in the window, the bypass gray value (mixed halftoned value as output by the blending processor 24) would be used instead. In this regard, it should be noted that the GRET adjustable threshold / detector 26 additionally provides a bypass for the GRET processor 28 at the output of the mixed arithmetic processor 24. Bypass so that the selection device 32 can select whether to pass GRET processed data modified by the GRET intensity selector 30 or to bypass data representing mixed halftoned data output by the blending arithmetic processor 24. In addition to the data, a selector signal is provided as an input to the GRET or bypass selection device 32. Thus, similar to the method and apparatus of the present invention, in addition to improving quality for unsaturated text, anti-aliasing for roughly saturated text / graphics is achieved.
[0012]
In FIG. 2, a method for calculating the contrast index is shown (which is used by the adaptive screen analyzer 14). In this method, a window of nine pixels (which is obtained from the output of the global color processing control device 12) is used, and between each adjacent pixel to determine the maximum difference between a pair of adjacent pixels. The absolute value of the difference is examined. In this regard, reference is made to US Pat. No. 5,956,157.
[0013]
FIG. 3 illustrates how the blending coefficients (BC1, BC2) are calculated by contrast index in two screener fuzzy logic systems. It further illustrates how the mixed halftone value (corrected value) is calculated from the mixing factor and the halftone dot gray value of the different halftone screen look-up table output.
[0014]
The system tunes the multi-level output in the LUT independent of the two halftone screens and the GRET LUT edge value. It is desirable to match the two halftone screen gray levels with the screen structure. In this regard, a density and structure match is formed in the boundary region between each image type, whereby the gray values in the two screener LUTs are adjusted to obtain this match. That is, at the same input value, the output densities of the two screens (these are not necessarily the same gray output value because each screen is different) are selected to match well (of course, the two screens Each screen structure is also selected to reduce texture mismatch). This makes it possible to achieve a gradual transition between each image type region. For similar reasons, each gray value and each GRET LUT (high / medium / low: high, medium) to give optimized performance for roughly saturated text anti-aliasing effects (depending on customer choice) , Low means a difference in strength with respect to anti-aliasing), but is adjusted independently of each screener LUT value. The system described here provides independent means for all that needs to be done.
[0015]
<Preferred color saturation adjustment>
After each image has already been RIP (raster image processing), even when the printer is in operation, the 1D (one-dimensional) overall color processing control LUT 12 is used first and the desired color at the last moment The possibility of adjustment is given. The input to the LUT 12 is 8-bit input data of a color separation image. The second input to the LUT 12 is a color tweaking value for adjusting the saturation of the color separation image. As shown in FIG. 22, the color saturation of the output by the operator performing gray level input into the LUT 12, the corresponding gray level output from the LUT 12, and the color fine adjustment input valid on the control panel of the workstation WS of FIG. A schematic diagram is provided showing the range of adjustments that can be made by modifying. It will be noted that this input follows the job image buffer and that the image data has been effectively modified after the image data has been output from the job image buffer. Thus, when the data is provided in electronic form, when making a copy (such as a proof copy) with various fine-tuning devices without rescanning the original hardcopy document and re-rasterizing the image data. In addition, an experiment may be performed by an operator. When viewing with the proof print, if the user does not like the printed color, the fine adjustment of the preferred color has a final step of minor color adjustment so that the user can adjust the color. Thus, a color with reduced saturation (de-saturated color) may be adjusted to a more saturated color. It may comprise increasing a specific color in the image. Prior to rasterization, we do not consider coloring for fine color adjustment to make the colors accurate or match colors, such as having a known color management process in the front end of the device . Prior to halftoning, preferably full color or process color because there is an improved result obtained by modifying the contone data rather than halftoned data. Color fine adjustments are made for the processes (cyan, magenta, yellow, and optionally black). The advantage of adjusting the contone data is that the dot structure formed after the halftone process or modifications to the dot data can introduce undesirable artifacts (interactions from other color channels) within the dot structure. And tend to give more color change, or at least tend to be more difficult to predict / control color adjustment.
[0016]
An adjustable GRET threshold stage is further provided in the threshold detector 26 for different degrees of roughly saturated text and graphics required by anti-aliasing to handle a somewhat limited degree of GCR / UCR range. It has been. Another improvement is the placement of one or more image screens in the screener 2 (screener 20) LUT so that different image screens can be selected in the printed page without having to reload the LUT ( Of course, the screen address calculator's screen positioning increment calculation needs to be changed from one image screen to another.) Further improvements include using two or more screens simultaneously in a blending operation for a smoother transition.
[0017]
FIG. 4 shows a detailed implementation of functions such as an adaptive screen analyzer 14 (which generates a contrast index; see FIG. 2 for a description of its function) and a mixing factor LUT 16 (for a description of its function). 3) and details of the blending operation block 24 (which uses the output value from each screener and the blending factor as a pointer to obtain the output value). ing. In this case (to obtain a very fast operation) to generate output (mixed halftone data; see FIG. 3 for this equation) for the GRET block (see FIG. 1 for details) A precomputed LUT approach is described. As can be seen in FIG. 3, after the contrast index is calculated, a blending factor is generated according to the description of FIG. For the contrast index example 0.4, the output value from screener 1 (18) is multiplied by 70 percent, while the output value from screener 2 (20) is multiplied by 30 percent. As can be appreciated from FIG. 3, the relatively small or relatively large contrast index is multiplied by 100 percent by one screener value and by 0 percent by the other screener value.
[0018]
FIG. 5 shows the detailed implementation of functions such as the screen address calculator 22, the screeners 18, 20 using the LUT (for high-speed computation), and the mixed computation block 24 (Brenda). In order to achieve higher speeds, a two-stage process channel approach is used. In this two-stage channel approach, the current even pixel and the current odd pixel are processed separately and simultaneously. To calculate the current even pixel contrast index, only each odd pixel adjacent to the current even pixel is required. For the current even pixel, a first-in first-out buffer (FIFO) 21a is provided to store each adjacent odd pixel required for the contrast index that defines the current even pixel. Similarly, a FIFO 21b is provided for the current odd pixel and stores each adjacent even pixel required to determine the contrast index for the current odd pixel. Each current even pixel is input to an even pixel screen LUT 18a, 20a, respectively; and each current odd pixel is input to an odd pixel screen LUT 18b, 20b, respectively. The output of each screener and the blending coefficient calculated from each contrast index for each of the odd and even pixels are input to the respective pixel blending arithmetic processors 24a and 24b. For each reasonable screen, the coordinate address of the modified screen value (based on the screen angle and screen frequency, different color separations can use different screener angles and frequencies, making these addresses different. Is generated according to the description below. In order to obtain the coordinate output for the halftone screener LUT (18a, 18b, 20a, 20b) that stores the halftone screen multi-stage output based on the input pixel value and the calculated coordinate value, the pixel clock and the line clock are rational. Used to increment the counter based on the current pixel position for a particular screen.
[0019]
<Generate modified screen value>
Referring to FIGS. 6, 7 and 8, each screen shows a regular screen tile formed in a 19 × 19 rectangular array. Screen tiles are 4x15 rotated rectangles. Screen tiles are used to modify 154.6 LPI (lines / inch) screen creases at 600 dpi (dots / inch) within a screener angle of 14.93 degrees. In each figure, it will be appreciated that the tile represents a halftone rendering value for a gray value of one of 255, 128, 2 in an 8-bit / pixel system.
[0020]
Referring to the data of FIGS. 9, 10, and 11, this can act as a repeatable sequence representing each halftone tile for each gray level plane 255, 128, 2 respectively. It represents 241 character strings. In FIG. 9, 241 numbers are seen in different rows and columns, but 241 numbers can be seen, such as a single row or brick consisting of 241 numbers. In the case of the plane 128 and the plane 2, it is more clear that not all the numbers in the brick are the same in the case of a normal gray level. As shown in FIG. 9, the brick width is 241, the brick height is 1, and there is a brick offset ID of 177 (this will be described below). The brick concept is used to show that 241 values corresponding to each gray level are used to define a halftone correction value for each pixel location in the image. It will of course be understood that these 241 values are determined based on the screen dominant frequency, the screen angle, and the size of the screen tile, and that these 241 values represent halftone correction values for a single color separation. Let's do it. In particular, when used to create the same multicolor image for an image halftone screen, it is usually desirable for each color separation to have a screen angle that is different from the screen angles of the other color separation colors.
[0021]
With reference to FIG. 12, a diagram is provided to aid understanding of the concept description using a brick sequence of the number of correction values. As shown in FIG. 12, the pixel value P (x, y) is input to an optional lookup table used when the input pixel is in an order different from the order of correction values (bit depth). Thus, if the input pixel has a gray level bit depth (eg, 12), the input pixel may be converted to a bit depth of 8 by a lookup table. The pixel to be halftoned and changed to be the proper bit depth is shown as g (x, y). The gray level of this input pixel is one of 256 brick planes 0-255 or acts as a pointer to one of 256 brick planes 0-255. Each brick plane contains a sequence of bricks for that gray level. Therefore, the plane 255 includes the 241 number sequences shown in FIG. In addition to the gray value of the pixel g (x, y), it also includes an x, y coordinate position or image pixel address for the image pixel. The coordinate position is used to place a unique correction value for the pixel in the pixel plane formed by the gray level for the pixel.
[0022]
Reference is made to the flowchart of FIG. 13 for calculating the coordinate values I, J in the brick plane (here, the coordinate values of the current pixel in the x, y image plane are known). In this embodiment, it is assumed that the coordinate value J is always equal to 1. This is because the brick height is 1 in this special case due to the nature of the halftone screen line. For other screens, the brick height may be 2 or more.
[0023]
To determine the correction value for pixel g (x, y), consider the first line of pixels where the gray level plane is determined by the gray value of the pixel and to be corrected. The coordinates of the image plane of the first pixel in the first line of each pixel are X = 0, Y = 0. The first number (I = 0, J = 0) in the brick of the gray level plane of the pixel is a correction value for the pixel. The second pixel (X = 1, Y = 0) in the first line of the image plane is modified by the second number in the brick of the brick plane having the gray level for the second pixel, followed by the 242nd The correction is made for each pixel g (x, 0) in the first line until the pixel is corrected. For this pixel, return to the beginning of the brick string or number sequence and repeat from the brick coordinates I = 0 to I = 240 in an iterative manner, and continue until all pixels for line Y = 0 are halftone corrected.
[0024]
For the next pixel line Y = 1, the first pixel in this line g (0,1) is mapped to the offset position I = 177 in the brick (this position is specific to this screen) And each different line of the image is found so that the starting position in the brick starts at a different calculated offset position. The next pixel in image line g (1,1) is mapped to rendering position I = 178 and continues until position 240 is reached. Next, the mapping of the next pixel in the image line is started from the rendering position I = 1. Thus, the offset is only used to start a new image line at various calculated offset positions. Thus, for pixels of the second image line Y = 1, the pattern is (from image pixel X = 0 to X = 63) I = 177 to I = 240, (image pixel X = 64 to X = 304). I = 0 to I = 240, (for image pixels X = 305 to X = 545) I = 0 to I = 240, etc. Mapping sequence until all pixels in this line are halftone corrected It is. For the following line Y = 2, the repetitive pattern is I = 113 to I = 240, I = 0 to I = 240, I = 0 to I = 240, etc., until all pixels on this line are halftone corrected. It is. It should be noted that for each pixel to be modified, the variable will be its gray level value so that a different brick plane is considered for one pixel with a pixel basis that depends on the pixel's gray value.
[0025]
The rough execution process of this process is shown by the flowchart of FIG. Here, a pixel having coordinates (x, y) is mapped to a certain position (I, J) in the brick plane. This position is supplied as one input to the halftone screen lookup table. The halftone screen lookup table also has an input to it (pixel gray value g (x, y)). The look-up table stores corrected pixel values for halftone correction of image pixel g (x, y). In this embodiment, there are 241 × 255 correction values in the LUT (brick width times the number of brick planes). In this embodiment, each pixel having a gray value of 0,255 is modified with the respective value, so that by assuming that the gray values of 0,255 have irrelevant I and J values, It is possible to further reduce the size. In the flowchart of FIG. 13, pixel image coordinate values x and y are input to the computer. The calculator takes the x-coordinate value and adds the y-coordinate value first divided by the brick height and then multiplied by the brick offset value to the x-coordinate value. This sum is then divided by the brick width. Here, the remainder is held as a brick coordinate value for I. For example, when X = 178, Y = 1, Bh = 1, Bs = 177, and Bw = 241, the sum is calculated to be 178+ (1/1) 177 = 355, which is the brick width 241 And the remainder I = 114 is generated. The J coordinate value is determined by taking the y coordinate value of the image plane, dividing the y coordinate value by the brick height, and holding the remainder as the J value. In the example for this screen, the value of J is always zero, but as mentioned above, some screens may have two or more brick heights, which The J coordinate is essential to determine. The brick coordinate calculator may be implemented by software, such as processed by a computer, or by a chip configured to perform this calculation. The calculation is expressed by equation (1).
[0026]
I = (X + (Y / Bh) * Bs)% Bw (1)
[0027]
Here, “%” indicates that division is performed and the remainder is determined. As described above, Bh in a certain situation is equal to 1, which simplifies the equation as in equation (2) in this case.
[0028]
I = (X + Y * Bs)% Bw (2)
[0029]
As can be seen in FIG. 5, simultaneous separation processing is performed for odd and even pixels, and hardware or software is executed to simultaneously calculate brick coordinate values for odd and even pixels. Also good. Further, since rendering is performed on both the halftone image screen and the halftone text screen, calculation of brick plane coordinates for the text screen and the image screen may be performed simultaneously. An example of a text screen is shown in FIG. 14, and a look-up table characterizing the correction values using Brick's planar technique to correct the pixels processed by the text screen is shown in FIG. . As can be seen from the figure, the text screen is much simpler than the image screen and does not require rotation between color separations as in the case of the image screen. However, the characterized unique text screen has two brick lines for each brick plane.
[0030]
Techniques for generating a modified halftone screen value lookup table are described with reference to FIGS. 26-31 and the flowchart of FIG. As is apparent, each step of the flowchart of FIG. 32 corresponds to each figure number of FIGS. In FIG. 26, a tile structure of an example screen having a screen angle of 45 degrees and 600 dots per inch and 141 lines per inch is shown with light and shade. Each pixel identified as C1 represents a pixel belonging to the same tile. It will be appreciated that all image planes have identical tiles joined together. In this example, each pixel forming a tile may form one cell or supercell, and each tile may be formed with multiple cells or multiple supercells within the tile structure. It will be understood. If the tile has a plurality of cells or supercells, there may be two sets of consecutive pixels in the tile.
[0031]
The individual pixels of this example tile have a unique position relative to other pixels in the tile and can be identified in this embodiment as pixels having a continuous number from 1 to 18. In general, the shape of the tile structure and the number of pixels in the tile structure and the direction of the tile are a function of the screen frequency and the screen angle. In FIG. 27, individual pixels within a tile are identified by a consecutive number from 1 to 18. In FIG. 28, the image plane is filled with the number of consecutive tiles. 29 and 30 show the results of a search for finding a continuous number of repetitive rectangular blocks in the image plane. As can be seen from the figure, it can be seen that the smallest repeating block or brick has six consecutive brick widths (Bw) and three consecutive brick heights (Bh). As can be further seen, the second course of each brick starts at a position offset by three consecutive numbers, which is referred to as brick offset or Bs.
[0032]
After defining the brick width, brick height, and brick offset parameters, the values for the modified value look-up table may be used instead of the number of consecutive pixels. For this particular screen, the number of runs for each pixel is consistent with all tile gray level values 1-255 for an 8-bit / pixel system. However, for each tile gray level value, the special sequence number of tiles corresponds to a specially modified value. FIG. 31 shows a value for a tile having a modified gray level value 106 while a pixel having a sequence number 1 for gray level 2 has a modified gray value 0 while all other pixels in the brick have a modified gray value 0. Is shown in In the example of a tile with gray level 128, it can be seen that some pixels in the tile have been modified to a value of 0 while other pixels have been modified to a pixel value different from 0. . At tile gray level 255, in this example, all pixels in the tile have a modified value 255.
[0033]
Referring to FIGS. 33 (a) -35 (b), there are shown different screen tile structures representing a screen structure of a tile having four cells or supercells in the tile structure. This tile structure corresponds to a screen having 171 lines per inch at a rotation angle of 0 degrees. As can be seen in FIG. 33 (a), the four cells have three different shapes. The brick structure of this tile is also shown in FIG. It can be seen that this brick structure has a brick height 7 with no brick offset. In FIGS. 34 (a), 34 (b), 35 (a), and 35 (b), the tile gray level 2,128 is modified for halftone dots each having a dispersed dot type growth pattern. Brick and tile structures with pixel values are shown. In a halftone dot growth pattern of gray level dots in this type of cell, the growth tends to be distributed to several pixel elements of the cell as the cell gray level increases. This growth pattern is different from the full dot type growth pattern. Here, the cell gray level is increased by increasing the gray level of one pixel until the pixel is at the maximum gray level (at this point the cell gray level growth tends to increase at the next pixel position in the cell). There is a tendency for growth to increase. Note also that the brick structure corresponds to the tile structure of this example.
[0034]
Various tile parameters such as screen angle, lines per inch, and number of gray levels per pixel are considered to generate a modified screen value for the tile. Furthermore, the characteristics of the dot driver and the dot type growth pattern are also considered. An example of a dot driver is shown in FIG. 36 (b) for a 16 × 16 dot size driver with a circular or spiral type growth pattern (where the dots in the cell tend to grow outward from the center). Has been. Other types of dot drivers may be used and are suitable for growth patterns of other shapes such as growth along lines or ellipses. Dot membership function where these factors take into account the cells in the tile and the contribution of exposure overflowing from other pixel locations of other cells that form part of the tile at the exposed pixel location It may be input to the generator. A screen profile builder may then be used to determine the overall gray level in the tile by summing the exposure values at each pixel location that has not yet been quantized. The screen profile quantizer then quantizes the individual pixel correction values so that the correction values can be expressed as integers (e.g., 0-255 for systems with 8 bits per pixel bit depth).
[0035]
Of course, as described herein, the assignment of a modified screen value means that the value is output directly to the printer when performing other image processing operations based on the obtained modified screen value. It will be understood that they do not. Thus, as described herein, the modified screen value for a particular pixel may be thresholded to establish eligibility for further processing such as edge enhancement processing or blending operations.
[0036]
An example functional block portion of an edge enhancement processing system that may be used in the method and apparatus of the present invention is shown in FIG. As already noted, the input to the GRET processor 28 is in the form of a binary bitmap after adjustment by the GRET adjustable threshold / detector 26. This relates to data subjected to threshold processing. Data output by the mixed arithmetic processor 24 is also bypassed to the GRET or bypass selection device 32. The input to GRET processor 28 is a binary bitmap. Here, the description of a “binary” bitmap or “binary” image means that the image pixels are fully or substantially fully exposed, or not exposed or substantially exposed (ie, It will be understood by those skilled in the art that it refers to a bitmap or image (where grayscale pixel data is substantially absent). In this example, since the GRET processor is processing pixels with a bit depth of 4 bits per pixel, the detector 26 has 8 bits per pixel image data, one pixel bit depth required by the GRET processor. It may be corrected to 4 bits per hit. The description “grayscale” refers to each pixel to indicate one or more shades of gray between fully exposed and not fully exposed. Indicates image data represented by data of 1 bit or more. Of course, the actual color of the pixel will depend on the color toner or pigment used in the printing process to develop the pixel. As an example in which image data is represented by four binary bit information, the binary bitmap has image data represented by 0 or 15. The binary bitmap has a matrix of this image data, where 0 represents unexposed pixels and 15 represents pixel areas that are fully exposed. Of course, a pixel area in which 15 represents an unexposed pixel and 0 is completely exposed can be used. The development location is preferably done in the exposed pixel area and not developed in the unexposed pixel area (known as discharge area development or reversal development, or charged area development may be used). Although reference has been made herein to “exposed” and “unexposed” pixels, the system may be used in other printing or display systems, even though the system characteristics do not use exposure (eg, inkjet using ink deposition). The pixels are displayed in accordance with the characteristics.
[0037]
In the GRET processor 28, the current pixel position is indicated by the description of n (i, j) as an output from the band buffer 100. The Sobel gradient masks 120 and 140 in the horizontal direction and the vertical direction act on the binary bitmap data n (i, j), so that the gradient x operator (gx) and the gradient y operator (gy) Is generated. Usable conventional Sobel gradient masks include the masks described in US Pat. No. 6,021,256, which is incorporated herein by reference. Other gradient masks may be used. The gradient magnitude or gradient magnitude (gm) is calculated by the processor 160 as the sum of the square of the gradient x operator (gx) and the square of the gradient y operator (gy) for each position in the bitmap. Calculated by taking the square root, a gradient magnitude map is generated. The gradient magnitude map is then stored in buffer 180 for later use. Similarly, a gradient angle (ga) 220 is determined for each pixel position, and a gradient angle map 220 is generated. For convenience, the gradient angle (ga) is preferably limited to the gradient direction (gd) selected by the gradient direction sorter 240. The gradient direction for each position is stored in buffer 260. The original bitmap data, and the gradient size (gm) and gradient direction (gd) corresponding to the data are supplied to the decision matrix 280. This matrix uses this information to select edge enhanced grayscale output data to replace the binary bitmap data that is input to the GRET processor. The decision matrix 280 compares the pixel data with a set of criteria represented by a predetermined pixel value and gradient magnitude, so that the central pixel of the binary bitmap data window is a black pixel or a white pixel. It is determined whether the pixel is a pixel and whether the central pixel is included in a pixel position corresponding to one pixel line and a defect site (kink site). In accordance with rules establishing a set of criteria, decision matrix 280 generates addresses that are provided to lookup table LUT 30. LUT 30 generates edge enhanced grayscale output data based on the address generated by decision matrix 280. The enhanced grayscale output data replaces the binary input data with the output by the threshold / detector 26, and the printer grayscale printhead (e.g., laser, LED, thermal, inkjet, or other type of printhead) ), Or when applied to a gray level display such as a CRT or other suitable display, it produces a smoother image without jagged edges. A computer that runs on hardware such as a general purpose computer, or a specially programmed computer, or a pipeline processing system (especially in the form of an application-specific integrated circuit (ASIC) or a combination of the ASICs). It will be appreciated that the GRET system can be implemented as a program. The LUTs 30 noted in FIG. 1 are a series of high / medium / low LUTs 30 that may be selected by input of a GRET intensity selector signal to provide a preference for edge enhancement intensity types.
[0038]
<Variable strength GRET>
Referring to FIGS. 17 and 18 to 21, the effect of the variable intensity of the GRET output is described. FIG. 18 shows 8 bits per pixel to indicate that the pixel area is binary and the value 255 represents maximum development, while the pixel area indicated by 0 is not developed or is background. The original image represented by is shown. The image is a line extending at different angles with respect to the origin and represents various lines originating from the origin position. Some of these diverging lines have a staircase effect or jaggedness, and minimize this jaggedness by placing gray level pixels at specific locations around the perimeter of each line so that they have a relatively smooth appearance. Note that it is the purpose of this resolution enhancement device to try to do. Considering FIG. 19 where the GRET output is shown, the look-up table 30 has been adjusted for intermediate intensity. When comparing FIG. 20 and FIG. 21 with FIG. 19, in the case of GRET output using a high-intensity look-up table, the gray level values given by the GRET processor are the cases of high intensity, intermediate intensity, and low intensity. Note that they are different. Note that values that are binary (ie, 0 or 255) are not affected. This therefore provides further adjustments to the operator of the workstation WS, as it allows the operator to adjust personal preference inputs for improvements in anti-aliasing. The operator simply selects which option of the LUT 30 (high intensity, medium intensity, low intensity) the user prefers to improve the jagged reduction.
[0039]
<Adjustable threshold input for GRET processing>
FIG. 23 shows a printer or display device 400 having the image processing system 10 described above. The device has a document scanned by a scanner 410, which generates an 8-bit signal representing the scanned density. Typically, raw image data scanned in red, green, blue (R, G, B) format is buffered in buffer 412 and then color processed and subjected to other image processing such as gamma correction 414. When the image data is in the format of a single color system, the color image data needs to be converted into a different color system by the color conversion calculation unit 416. The converted color separation image data normally used in a printer is preferably C, Y, M, K. As described above, the color conversion processor may include well-known undercolor removal and / or gray component replacement functions. The undercolor removal function mainly removes chromatic colors (yellow, magenta, cyan) in dark areas or roughly intermediate shade areas to reduce toner height or toner coverage. It is. Gray component replacement is similar to this, but refers to the use of black toner for multi-colored gray components and is not limited to undercolor removal in roughly neutral color areas. Although the purpose of these two techniques is different, in practice it is similar in that black toner is used to reduce some of the colored toner from the image. Referring to FIGS. 24 and 25, an example of GCR and UCR having a brown mixed color is provided. The GCR function allows the gray component of colored printing ink or toner to be replaced with a black process color due to effects in the overall color space. The replacement amount can be set to a desired amount. The color impression remains the same. No color is required to produce a particular shade (ie, area is reduced). This means that the gray axis is more stable. Since fewer chromatic colors are used, the cost can be reduced. UCR is a setting option for addition or selection in color reproduction. In this step, the gray component of the chromatic printing ink or toner is replaced with black in the shade of the intermediate image. Less color is needed to produce a particular shade (ie, area is reduced). This means that the gray axis is more stable, less chromatic colors are used, and the cost can be reduced by UCR. Although it is known to have a UCR and / or GCR processor after color space conversion, it is more preferable to have a UCR and / or GCR processor during color space conversion. The problem with using UCR and / or GCR is that the most saturated color value produced by the process does not reach a level that would otherwise represent binary data image information. For example, the threshold / detector 26 has a preprogrammed threshold level value that is higher than the threshold level assumed to be binary information. If a color conversion process is used and all processing information falls below the preprogrammed threshold level, it is assumed that all the information is not a binary image data file and the GRET processor Will be sent to the bypass. The operator of the printer knows the color conversion process used and is therefore not useful above the level indicating a binary image data file by adjusting the threshold input to the GRET adjustable threshold / detector 26. It is possible to provide a new threshold level considering whether the level is substantially regarded as a threshold value. For example, typically a binary image file is represented by a saturated color gray value in an 8-bit depth system (where the gray value is considered to be 254 or 255). Therefore, the threshold value 253 is set by the threshold value / detector 26. However, the maximum gray value will not be greater than 253, especially if UCR and / or GCR are used. Therefore, it should be noted that no binary image data file will be generated and only data that bypasses the GRET processor will be selected. However, this result does not match the characteristics of the image information due to the processing characteristics in color conversion. In order to solve this problem, the operator can decide what is the binary image data file so that improved control is performed by the image data selected from either the GRET processor bypassed data or the GRET processed image data. You are given the opportunity to enter a programmable adjustable threshold to set a new threshold to determine what should be. Thus, for example, when UCR and / or GCR are used, the operator may use a GRET threshold to ensure that some information selected for output is GRET processed information. / Set a lower threshold for detector 26 (eg 253). Alternatively, a lower threshold can be obtained by automatically changing the threshold by selection by an operator of undercolor removal and / or gray component replacement, or by adjusting the amount of undercolor removal and / or gray component replacement. May be set.
[0040]
Other modifications known to the prior art may be made on the scanned raw image data. Further, the input from the electronic data source 420 may be supplied to each page of image data that is further input to the job image buffer 424 after being rasterized by the raster image processor (RIP) 422. One or more pages of rasterized image data from a scanner or electronic data source are stored (preferably in compressed form) in a job image buffer (JIB). Thus, after the data is transmitted to the printer, the image data in the job image buffer is electronically recirculated, so that a plurality of documents arranged in page order can be printed. In this regard, see US Pat. No. 5,047,955 filed in the name of Shope et al. The present invention includes the contents described in the above specification. Image data is output to the image processing system 10 described above for final output to a gray level printhead or display 470. The printhead may be modified by a writer interface board 460 for modifying for non-identical recording elements, or other known modification devices or schemes such as adjusting the exposure level by pulse width modulation, pulse intensity modulation, etc. . In this regard, see US Pat. No. 6,021,256 filed in the name of Ng et al. And US Pat. No. 5,914,744 filed in the name of Ng. Overall control of the device may be provided by a marking engine controller 426 in the form of one or more microcomputers suitably programmed for control by well-known programming skills. Workstation WS includes GRET adjustable threshold input values used by detector 26, GRET intensity selection (high LUT, medium LUT, low LUT), and real time color fine adjustment used in LUT 12; Various job parameters relating to the print job, such as the number of copies and paper selection, are input to the marking engine controller.
[0041]
In a preferred apparatus of the present invention, a printhead (eg, 600 dpi resolution) exposes a uniformly charged photoconductive drum or web and the web develops with pigmented electroscopic toner particles. The image appears. The development and other color separation developments are then transferred to the receiving sheet, either continuously in a separation step or in one step, either directly from the photoconductive web or drum or indirectly via an intermediate transfer member. . In this regard, reference is made to US Pat. No. 6,075,965, which is given in the name of Tombs et al. And describes a color electrophotographic apparatus for the continuous transfer of each color separation image to a receiving sheet.
[0042]
An extension of the method includes storing one or more image screens contained in one screener within each screener, thereby allowing each different image screen to be printed without reloading the screen page (or screener LUT). It can be used for the next page). Of course, in this case it would be necessary to store more than one set of screen addresses (in each LUT row and column). Furthermore, an image screen selector function for use would also need to be included, for example in the workstation WS.
[0043]
For other extensions of the spirit of the present invention, this method with a lower addressable output device can be used to use unreasonable screens and get more choices for precise screen angles and frequencies. It is possible that an irrational screen coordinate calculator (errors for screen angle and frequency calculations can be propagated in the process flow direction so that subsequent screen block adjustments correct these errors. Can be used). More specifically, the screen coordinate calculator calculates the LUT data address for each stage via the screen brick and accumulates position errors due to progressing through the brick. This position error is corrected by jumping the address when a predefined position error threshold is exceeded.
[0044]
Therefore, image data that has undergone conventional processing such that the gray level value does not substantially match image data as binary image data, and thus such image data is free from processing according to anti-aliasing processing. An improved apparatus and method has been described that provides tunability in the process for. With this adjustability, the image data can be properly matched and processed by edge enhancement processing.
[0045]
The invention has been described with reference to the preferred embodiments. However, it will be appreciated that variations and modifications can be effected within the scope of the appended claims.
[Brief description of the drawings]
FIG. 1 is a schematic block diagram of an image processing system according to the present invention.
FIG. 2 is a diagram showing a window of 9 pixels, showing one specific method for obtaining a contrast index.
FIG. 3 is a diagram illustrating determining blended and modified dot values using a graph showing a blending factor versus contrast index (for two different halftone processes).
FIG. 4 is a block diagram showing an enlarged part of FIG. 1 in detail.
FIG. 5 is another block diagram showing an enlarged detail of a portion of the system of FIG. 1;
FIG. 6 illustrates a 19 × 19 pixel halftone screen tile that may be used as one of each halftone screen in the system of FIG. 1;
FIG. 7 illustrates a 19 × 19 pixel halftone screen tile that may be used as one of each halftone screen in the system of FIG. 1;
FIG. 8 shows a 19 × 19 pixel halftone screen tile that may be used as one of each halftone screen in the system of FIG. 1;
FIG. 9 illustrates an example of a screen address “brick” used to generate a halftone screen pixel correction value for the screen illustrated in FIG. 6;
10 illustrates an example of a screen address “brick” used to generate a halftone screen pixel correction value for the screen illustrated in FIG. 7. FIG.
11 is a diagram illustrating an example of a screen address “brick” used to generate a halftone screen pixel correction value for the screen illustrated in FIG. 8. FIG.
12 is a diagram illustrating examples of look-up tables used to generate halftone screen pixel correction values for the illustrated screens of FIGS. 6, 7, and 8. FIG.
FIG. 13 is a flow diagram that may be used to determine the address to which a brick is associated in the lookup table structure of FIG.
FIG. 14 illustrates an example of a text screen tile used to define a halftone modified text screen value.
FIG. 15 illustrates an example of a screen address brick used to generate a halftone screen pixel correction value for the text screen of FIG.
FIG. 16 is a block diagram of a preferred gray level edge enhancement processor used in the system of FIG.
FIG. 17 illustrates the interrelationships for each type of output from a gray level edge enhancement processor.
FIG. 18 schematically illustrates a binary image where 255 represents the maximum density and 0 represents the background or no density.
FIG. 19 schematically illustrates a binary image with gray level edge enhancement as a function of output from an intermediate intensity setting for a lookup table.
FIG. 20 schematically illustrates a binary image with gray level edge enhancement in response to output from a low intensity setting for a lookup table.
FIG. 21 schematically illustrates a binary image with gray level edge enhancement in response to output from a high intensity setting for a lookup table.
FIG. 22 is a graph showing a relationship between an input pixel gray value and a corrected gray value according to color saturation fine adjustment.
FIG. 23 is a block diagram of a printing or display system incorporating the image processing system of FIG.
FIG. 24 is a diagram illustrating an example of gray component replacement (GCR) in a color conversion process.
FIG. 25 is a diagram illustrating an example of under color removal (UCR) in a color conversion process.
FIG. 26 shows steps for forming a brick structure.
FIG. 27 shows steps for forming a brick structure.
FIG. 28 is a diagram illustrating steps for forming a brick structure.
FIG. 29 is a diagram illustrating steps for forming a brick structure.
FIG. 30 is a diagram illustrating steps for forming a brick structure.
FIG. 31 is a diagram illustrating steps for forming a brick structure.
FIG. 32 is a flowchart of a process for forming a brick structure.
33A shows a tile structure for a screen having 171 lines per inch at a rotation angle of 0 °, and FIG. 33B shows the brick structure.
34A shows a tile structure for a screen having 171 lines per inch at a rotation angle of 0 degrees, and FIG. 34B shows the brick structure.
FIG. 35A shows a tile structure for a screen having 171 lines per inch at a rotation angle of 0 °, and FIG. 35B shows the brick structure.
FIGS. 36A and 36B are diagrams showing dot size drivers that are used when generating a modified screen value for one tile having a growth pattern of a circular or spiral type. FIG. .
[Explanation of symbols]
26 Adjustable threshold / detector (adjustable threshold device)

Claims (3)

An edge enhancement processing system for modifying image data at a particular pixel location to include grayscale image data to reduce jaggedness in the image, for a current input gray level pixel according to a threshold criterion An adjustable threshold device for determining a current binary pixel value; an operator with input access to the threshold device for adjusting a threshold within the threshold criteria; and a jagged edge of the image An edge-enhanced image processing device that examines a current binary pixel and an adjacent binary pixel according to a predetermined criterion for determining to modify the current input gray level pixel to a gray scale value to reduce; With
Before Symbol image data of two screens and which Halftone reduction; over each mixing coefficient obtained from the contrast index of all screens to each halftone correction value of the screen, mixing the Hung halftone correction value Therefore, each pixel of the image data blender and for modifying said current input gray level pixel; computing said contrast index as the maximum value of the absolute difference between adjacent pixels for each pixel of the previous SL image data An edge-enhancement process comprising : an adaptive screen analysis device; and a look-up table used to calculate the mixing factor such that output densities of the screen match in accordance with the contrast index system. Determining an adjustable threshold within a threshold criterion in response to input from an operator; and determining a current binary pixel value for a current input gray level pixel according to the threshold criterion using a threshold. Determining; adjoining the current binary pixel and the pixel according to a predetermined criterion for determining to modify the current binary pixel to a grayscale value to reduce jagged edges of the image Inspecting each pixel; using gray scale values instead of the current binary pixels to reduce jagged edges of the image;
Prior to the determination of the adjustable threshold, a contrast index is calculated as a maximum value of absolute differences between adjacent pixels for each pixel of the image data, and two screens are calculated according to the contrast index. calculate the mixing coefficients as output density coincide, the image data is halftoned with the two screens, over the respective mixing coefficients obtained from the contrast index of all screens to each halftone correction value of the screen An edge enhancement method for processing image data, further comprising calculating the current input gray level pixel by mixing the multiplied halftone correction values. Processing image data using undercolor removal and / or gray component replacement; whether or not undercolor removal and / or gray component replacement is used, or the range of such use, said image data Adjusting the edge enhancement process of
The step of adjusting the edge enhancement processing of the image data includes:
Halftoning the image data on two screens;
Multiplying each halftone correction value of the screen by each mixing factor obtained from the contrast index of the entire screen;
Mixing the multiplied halftone correction values;
Determining a current binary pixel value for each pixel of the mixed image according to whether undercolor removal or gray component replacement is used, or the extent of such use;
Inspect the current binary pixel and each pixel adjacent to the pixel according to a predetermined criterion to determine to modify the current binary pixel to a grayscale value to reduce jagged edges of the image And the stage of
Modifying the current binary pixel to the determined grayscale value to reduce jagged edges of the image;
The contrast index is calculated as the maximum value of absolute differences between adjacent pixels for each pixel of the image data,
The edge enhancement method for processing image data, wherein the mixing coefficient is calculated so that output densities of the screens coincide with each other according to the contrast index.

Images